GitHub Copilot Statistics 2026
By Axis Intelligence Research
Co-authors: Sarah Mitchell | Last updated: June 22, 2026 | Next scheduled update: Q4 2026 (October) | License: CC BY 4.
Quick Answer
GitHub Copilot had 4.7 million paid subscribers as of January 28, 2026, up 75% year-over-year, per Microsoft’s FY26 Q2 earnings disclosure. Total users reached 20 million by July 2025. The tool generates 46% of all code written by active users and is deployed at 90% of Fortune 100 companies. Axis Intelligence Research estimates GitHub Copilot ARR at $900 million to $1.1 billion as of Q2 FY2026 — a figure Microsoft has never disclosed directly.
Key Findings
According to Axis Intelligence Research analysis of Microsoft SEC EDGAR filings, GitHub official disclosures, and the JetBrains AI Pulse survey (n=10,000+, January 2026):
- GitHub Copilot reached 4.7 million paid subscribers by January 2026, up from 1.8 million in FY2024 and just over 1 million in early 2024 — a 161% increase over two years, per Microsoft’s public earnings commentary and SEC filings.
- The Copilot Platform Momentum Index (CPMI™), calculated by Axis Intelligence Research, rose from 4.8 in FY2024 to 34.5 in Q2 FY2026, revealing a growing structural tension: enterprise penetration and code-share metrics are accelerating, while developer satisfaction has collapsed from roughly 30% “most loved” to 9% — flagging a retention risk that user counts alone do not capture.
- GitHub Copilot generates 46% of all code written by active users as of 2025–2026, up from 27% at commercial launch in 2022, per GitHub’s own telemetry — with Java developers reaching 61% and the overall suggestion acceptance rate holding at 27–30%.
- Axis Intelligence Research estimates GitHub Copilot ARR at $900 million to $1.1 billion as of Q2 FY2026, computed from the disclosed 4.7 million paid subscriber base and public pricing tiers — above the Financial Times’ conservative floor of “at least $550 million” based on the cheapest tier only, and a fraction of GitHub’s total ARR of more than $3 billion reported at end 2025.
- On June 1, 2026, GitHub transitioned all Copilot plans to usage-based billing, replacing premium request units with AI Credits priced at $0.01 per credit and tied to token consumption — the most significant pricing restructure since the product launched in 2022, introduced precisely because agentic workflows consume compute that flat-rate pricing could no longer absorb.
GitHub Copilot User Statistics
How Many Users Does GitHub Copilot Have?
The distinction between “users” and “paid subscribers” matters here more than in most software products. GitHub Copilot has always had a large free and trial population — students, open-source maintainers, individual developers on the free tier — that inflates the total number while the revenue-generating base is much smaller.
| Metric | Value | Date | Source |
|---|---|---|---|
| Total users (cumulative) | ~20 million | July 2025 | Microsoft Q4 FY2025 earnings call |
| Total users (cumulative) | ~15 million | April 2025 | Microsoft Q3 FY2025 earnings call |
| Total users added in 3 months | 5 million | Apr–Jul 2025 | Microsoft earnings commentary |
| Paid subscribers | 1.8 million | FY2024 | Microsoft SEC EDGAR 10-K |
| Paid subscribers | ~4.7 million | January 2026 | Microsoft FY26 Q2 earnings call |
| Paid subscriber growth YoY | +75% | Jan 2026 | Microsoft FY26 Q2 earnings call |
| Pro+ subscription growth QoQ | +77% | Q2 FY2026 | Microsoft earnings commentary |
| Enterprise customers | ~77,000 | FY2024 | Microsoft SEC EDGAR |
| Enterprise customer growth | +75% QoQ | Q2 FY2025 | GitHub/Microsoft |
| Organizations using Copilot | 50,000+ | 2025 | GitHub official disclosures |
Microsoft’s FY26 Q2 earnings release, filed January 28, 2026, is where the 4.7 million paid subscriber figure originates. Microsoft also switched at this point from reporting “users” to reporting “paid subscribers” — a deliberate shift that signals the company now wants investors tracking the monetized base, not the broader user population.
The gap between 20 million total users and 4.7 million paid subscribers represents 15.3 million free users, students, open-source contributors, and trial accounts. At GitHub‘s total registered base of over 150 million developers, paid penetration sits at roughly 3.1% — which is simultaneously the largest paid user base of any AI coding tool and a signal of how much runway remains before Copilot approaches saturation.
GitHub Copilot Paid Subscriber Growth Timeline
| Period | Paid Subscribers | Source |
|---|---|---|
| Technical preview (2021) | ~0 (invite-only) | GitHub Blog |
| Commercial launch (Jun 2022) | Launched at $10/mo | GitHub Blog |
| Early 2024 | ~1 million | Microsoft |
| FY2024 (Jun 2024) | 1.8 million | Microsoft SEC EDGAR 10-K |
| Q3 FY2025 (Apr 2025) | ~3 million est. | Axis Intelligence Research est. |
| Q4 FY2025 (Jul 2025) | ~3.5 million est. | Axis Intelligence Research est. |
| Q2 FY2026 (Jan 2026) | 4.7 million | Microsoft earnings call |
| Q1 FY2027 (Jun 2026) | ~5.8 million est. | Axis Intelligence Research est. |
Satya Nadella confirmed in the July 2025 earnings call that GitHub Copilot had crossed 20 million total users — and noted that Copilot now represents a larger business than GitHub itself was at the time of Microsoft’s $7.5 billion acquisition in 2018. That comparison is worth anchoring: not a metaphor for scale, but a literal statement that a product that didn’t exist four years ago generates more revenue than the entire business Microsoft purchased for $7.5 billion.
GitHub Copilot Revenue Statistics
What Is GitHub Copilot’s ARR?
Microsoft does not disclose GitHub Copilot’s annual recurring revenue separately. What exists publicly is: the paid subscriber count, the public pricing tiers, and two external estimates.
Axis Intelligence Research ARR Model (Q2 FY2026):
Using the disclosed 4.7 million paid subscriber base and a conservative mix assumption (60% Individual at $10/month, 25% Business at $19/month, 15% Enterprise at $39/month):
| Tier | Subscribers | Monthly Price | Annual ARR |
|---|---|---|---|
| Individual (Pro/Pro+) | ~2.82M | $10–$39/mo (blended $12) | ~$406M |
| Business | ~1.175M | $19/mo | ~$268M |
| Enterprise | ~0.705M | $39/mo | ~$330M |
| Total (conservative) | 4.7M | ~$900M–$1.1B |
The Financial Times reported in May 2026, citing a person with knowledge of the business, that GitHub Copilot was on track to generate “at least $550 million annually” — a floor based on the cheapest subscription tier only, not the blended mix. Axis Intelligence Research’s calculation uses the actual pricing tiers and produces a higher range: $900 million to $1.1 billion, depending on enterprise discount assumptions.
The same FT report noted that GitHub’s total ARR exceeded $3 billion at end 2025. GitHub Copilot, on these estimates, accounts for roughly 30% of GitHub’s total ARR — the fastest-growing component, and the one responsible for most of GitHub’s acceleration since 2023.
Per Microsoft’s earnings commentary, Satya Nadella stated that Copilot’s revenue had grown larger than “GitHub itself” at acquisition in 2018 — implying an ARR above $200 million even by the conservative reading of that statement. The $900M–$1.1B estimate from Axis Intelligence Research is consistent with the product’s growth trajectory.
Copilot Platform Momentum Index (CPMI™)
Measuring Whether Copilot’s Enterprise Strength Offsets Satisfaction Erosion
The Copilot Platform Momentum Index (CPMI™) is an original composite metric calculated by Axis Intelligence Research. It measures whether GitHub Copilot’s raw growth metrics — subscriber growth, enterprise penetration, and code generation share — are sufficient to offset deteriorating developer satisfaction, which is the leading indicator of long-term churn.
CPMI™ Formula:
CPMI™ = (Paid_Subscriber_Growth_YoY% × Enterprise_Penetration% × Code_Share%)
÷ (Developer_Satisfaction_Score × 1,000)
Attribution: CPMI™ is a trademark of Axis Intelligence Research. Formula disclosed under CC BY 4.0. To cite: “Axis Intelligence Research CPMI™, calculated from Microsoft SEC EDGAR filings and JetBrains AI Pulse 2026, June 2026.”
| Period | Sub Growth YoY | Enterprise % | Code Share | Satisfaction | CPMI™ | Signal |
|---|---|---|---|---|---|---|
| FY2024 (Jun 2024) | +133% | 40% | 27% | 30% “most loved” | 4.8 | Growth dominant; satisfaction healthy |
| H1 FY2025 (Jan 2025) | +80% | 75% | 40% | 25% “most loved” | 9.6 | Enterprise expanding; satisfaction dipping |
| Q2 FY2026 (Jan 2026) | +75% | 90% | 46% | 9% “most loved” | 34.5 | Critical zone: enterprise anchors growth |
| Q1 FY2027 (Jun 2026 est.) | +65% | 92% | 46% | 9% “most loved” | 30.6 | Slight deceleration; satisfaction floor not yet found |
What the CPMI™ reveals: the index rises sharply from 4.8 to 34.5 not because Copilot is improving on every dimension, but because the denominator — satisfaction — has collapsed. A higher CPMI™ in this specific case is not good news. It is a warning signal: Copilot’s growth is now almost entirely driven by enterprise procurement and GitHub ecosystem lock-in, not by developers choosing it on merit against alternatives. When satisfaction sits at 9% “most loved” — versus Claude Code at 46% in the same JetBrains survey — and the CPMI™ keeps rising, it means the enterprise anchor is working, but it is working harder than ever just to compensate for ground being lost on the product experience side.
No metric surfaces this dynamic cleanly in existing public reporting. The CPMI™ was built specifically to make it visible.
GitHub Copilot Market Share Statistics
What Is GitHub Copilot’s Market Share in AI Coding Tools?
GitHub Copilot remains the largest AI coding assistant by paid user base and by developer awareness, but its share of developer preference — the metric that predicts future adoption — has narrowed materially since mid-2025.
| Tool | Paid Users / ARR | Work Adoption (Jan 2026) | “Most Loved” (JetBrains) | Source |
|---|---|---|---|---|
| GitHub Copilot | 4.7M paid (Jan 2026) | 29% of developers | 9% | JetBrains AI Pulse Jan 2026 |
| Cursor | $2B ARR (Feb 2026) | 18% of developers | 19% | Bloomberg; JetBrains |
| Claude Code | $2.5B ARR (Feb 2026) | 18% of developers | 46% | Anthropic; JetBrains |
| JetBrains AI / Junie | n/a | 11% of developers | n/a | JetBrains AI Pulse Jan 2026 |
| Amazon Q Developer | n/a | ~5% (est.) | n/a | Industry estimates |
Per the JetBrains AI Pulse survey published April 2026 (10,000+ developers across eight languages), GitHub Copilot has 76% awareness and 29% adoption at work — the highest of any AI coding tool on both metrics. But growth on both has stalled since the prior wave. Cursor and Claude Code have both reached 18% work adoption, matching each other and closing in on Copilot’s 29%.
The satisfaction gap is the number that matters most for forecasting. Copilot’s 9% “most loved” versus Claude Code’s 46% is a 5.1× differential. Among developers with more than ten years of professional experience — who tend to be the decision-influencers in enterprise tool selection — that gap is reportedly larger still. In the 10,000+ employee enterprise segment, Copilot’s adoption at 40% is structural: procurement relationships, GitHub Enterprise contracts, and Microsoft 365 bundling keep it deployed even when individual developers would choose something else.
GitHub Copilot Market Share: Volume vs Preference
The two things are genuinely different in 2026, and it is worth being exact about what each number means.
Copilot’s “42% market share” figure, cited in several reports, refers to its share among paid AI coding tool users — not total developer population. Among the roughly 20 million developers who have ever used any paid AI coding tool, Copilot claims about 42%. That number reflects distribution advantage (VS Code, GitHub, Microsoft enterprise relationships) far more than product preference.
Among the broader developer population — the roughly 85% of developers who now use at least one AI coding tool, per the Stack Overflow Developer Survey — Copilot’s 29% work adoption puts it first by usage, while its 9% satisfaction rating puts it last among the three major tools. Both numbers are real. The tension between them is what the CPMI™ is designed to measure.
GitHub Copilot Productivity Statistics
Does GitHub Copilot Actually Make Developers Faster?
The productivity evidence for Copilot is more complicated than either the marketing claims or the skeptic backlash suggest. There are at minimum four distinct bodies of evidence, with different populations, methodologies, and conclusions.
| Study | Methodology | Population | Finding | Source |
|---|---|---|---|---|
| GitHub/MIT controlled experiment (Peng et al. 2023) | Randomized controlled trial | 89 developers, JavaScript task | +55.8% task completion speed | Peng et al. 2023, confirmed arXiv:2302.07080 |
| Microsoft/MIT/Princeton/Wharton RCT (2024) | 3 separate RCTs | 4,867 developers at Microsoft, Accenture, Fortune 100 firm | +26.1% pull requests completed/week | InfoQ, September 2024 |
| Accenture internal study | Enterprise deployment | ~450 developers, then 12,000+ | 84% increase in successful builds; PR time 9.6d → 2.4d | Accenture/GitHub case study |
| METR independent study (2025) | Experienced developers on real tasks | Expert developers, real-world tasks | –19% task completion speed | METR, 2025 |
The four studies are not contradictory — they are measuring different things. The 55.8% speedup applies to a well-scoped JavaScript task completed by developers who could be randomly assigned to a tool they hadn’t used yet. The 26.1% increase in pull requests measures a production outcome across 4,867 developers in enterprise settings over time. The METR –19% result applies to experienced developers on complex, real-world tasks — the kind of work where AI tools offer the least autocomplete value and the most context-loading friction.
Reading the METR finding as “Copilot makes developers slower” would be as misleading as reading GitHub’s 55.8% as “Copilot makes everyone 55.8% faster.” The honest summary: measurable productivity gains exist on well-scoped tasks and for junior developers; experienced developers on complex work may see neutral or negative effects; production metrics like PR throughput improve for most enterprise deployments, but the magnitude depends heavily on task type and adoption pattern.
Microsoft Research’s additional finding — that it takes approximately 11 weeks for developers to fully realize productivity gains — is among the most practically important and least cited statistics in this space. Organizations that evaluate Copilot in a one-week pilot are measuring almost nothing.
GitHub Copilot Code Quality Statistics
| Metric | Value | Source |
|---|---|---|
| Average code share (% of active users’ code) | 46% overall; 61% Java | GitHub official telemetry, 2025–2026 |
| Code share at launch (2022) | 27% | GitHub disclosures |
| Suggestion acceptance rate | 27–30% | GitHub telemetry; Accenture study |
| Accepted code retention (characters kept) | 88% | Accenture/GitHub case study |
| Code readability improvement | +3.62% | GitHub research, November 2024 |
| Code reliability improvement | +2.94% | GitHub research, November 2024 |
| Lines without readability errors | +13.6% | GitHub research, November 2024 |
| Python code with potential security weaknesses | 29.1% | GitHub research analysis |
| Copilot-generated code with security issues (Pearce et al.) | 44% of 89 scenarios tested | Pearce et al. 2022, arXiv ICSE 2025 |
| Pull request time (Accenture) | 9.6 days → 2.4 days (–75%) | Accenture/GitHub |
| Successful builds increase (Accenture) | +84% | Accenture/GitHub |
| Code churn increase linked to AI tools | From 3.1% (2020) to 5.7% (2024) | GitClear, 211M lines analyzed |
The security numbers are the ones that get buried most often in coverage focused on speed gains. A 2022 study (Pearce et al., referenced in the 2025 ICSE proceedings) found security issues in 44% of Copilot-generated code scenarios across 89 cases covering 25 CWE categories. GitHub has improved its filtering since 2022, but the pattern is structural: models trained on public code absorb the vulnerability patterns present in public code. The 29.1% Python security-weakness rate in GitHub’s own 2024 research is the company acknowledging the problem exists.
The GitClear code churn finding — from 3.1% to 5.7% over four years, coinciding with AI tool adoption — is the most underreported statistic in the AI coding space. Higher churn means more code committed, then revised or deleted, which means AI is producing code that needs more rework, not less. It doesn’t invalidate the productivity gains on individual tasks; it suggests the gains at task level don’t always translate to cleaner codebases at scale.
GitHub Copilot Enterprise Statistics
GitHub Copilot Enterprise Adoption Data
| Metric | Value | Date | Source |
|---|---|---|---|
| Fortune 100 companies using Copilot | 90% | 2025 | GitHub/Microsoft official |
| Fortune 500 companies using Copilot | 60%+ operate 10,000+ seats | April 2026 | Microsoft earnings commentary |
| Enterprise customers (FY2024) | ~77,000 | Jun 2024 | Microsoft SEC EDGAR |
| Enterprise customer growth QoQ | +75% | Q2 FY2025 | Microsoft earnings commentary |
| Accenture seats deployed | 740,000+ | 2026 | Microsoft earnings call |
| Siemens deployment | 30,000 developers | 2025–2026 | Reported case studies |
| AMD monthly lines accepted | Hundreds of thousands | 2025–2026 | Reported case studies |
| Roche seats | 90,000+ | April 2026 | Microsoft earnings commentary |
| Developers installing Copilot on day 1 of access | 81.4% | Accenture study | GitHub/Accenture |
| Developers using Copilot 5+ days/week | 67% | Accenture study | GitHub/Accenture |
| License utilization when made available | 80% | Opsera analysis | Opsera, 2025 |
| New GitHub developers using Copilot in first week | 80% | GitHub Universe 2025 | GitHub Blog |
The Accenture deployment is worth dwelling on specifically. 740,000 Copilot seats across a single organization is not a pilot — it is the largest known single-organization AI coding deployment in history. That it was disclosed in Microsoft’s earnings call rather than Accenture’s own communications tells you something about whose marketing it serves. The productivity improvement Accenture cites — including the 84% increase in successful builds — comes from their own internal measurement, which is methodologically separate from the GitHub-commissioned studies and therefore more credible as an indicator of what enterprise deployment actually produces at scale.
GitHub Copilot Pricing Statistics 2026
GitHub Copilot Plans and Pricing (Current as of June 2026)
GitHub Copilot moved to usage-based billing on June 1, 2026. The official GitHub pricing page confirms the following tiers, unchanged in base price but now metered by AI Credits instead of premium request units.
| Plan | Monthly Price | AI Credits Included | Code Completions | Models Available | Target User |
|---|---|---|---|---|---|
| Free | $0 | 50 (limited) | 2,000/month | Select models | Individuals, casual use |
| Pro | $10 | $10 in credits (~1,000) | Unlimited | GPT-4.1, Claude, Gemini | Individual developers |
| Pro+ | $39 | $39 in credits (~3,900) | Unlimited | All models incl. Claude Opus 4.6, o3 | Heavy users, agents |
| Business | $19/user | $19/user in credits (~1,900) | Unlimited | GPT-4.1, pooled across org | Teams, SMBs |
| Enterprise | $39/user | $39/user in credits (~3,900) | Unlimited | All models, custom fine-tuning | Large orgs (GitHub Enterprise Cloud required) |
Per GitHub’s official pricing documentation, overage beyond the included credits is billed at $0.01 per AI Credit (1 credit = $0.01 USD), with model-specific per-token rates. Code completions and Next Edit Suggestions remain unlimited on all paid plans and do not consume AI Credits.
The Enterprise tier carries a frequently missed cost: it requires GitHub Enterprise Cloud at $21 per user per month. The real Enterprise price is therefore $60 per user per month, not $39 — a meaningful difference at large seat counts.
The April 27, 2026 GitHub Blog announcement framed the shift plainly: “Copilot is not the same product it was a year ago. It now powers far more complex, agentic workflows that consume far more compute. A quick chat question and a multi-hour autonomous coding session can currently cost the user the same amount, and that model is no longer sustainable.” The developer community’s response was immediate and critical — GitHub Community discussions accumulated hundreds of comments within days, with users reporting that a single development day’s agent usage could exhaust an entire month’s credit allotment.
Historical Pricing Changes
| Date | Change | Source |
|---|---|---|
| Jun 2022 | Commercial launch: $10/mo Individual, $19/mo Business | GitHub Blog |
| Dec 2024 | Free tier launched (2,000 completions, 50 premium requests) | GitHub Blog |
| Early 2026 | Pro+ adds Claude Opus 4.6 and o3 model access | GitHub pricing page |
| Apr 22, 2026 | New self-serve Business sign-ups temporarily paused | GitHub Docs |
| Jun 1, 2026 | Full transition to usage-based (AI Credits) billing | GitHub Blog |
GitHub Copilot Feature Statistics 2026
From Autocomplete to Agentic Platform
GitHub Copilot in mid-2026 is a different product category from its 2022 launch. The product has expanded from a single-purpose inline code completion tool to a multi-layered platform spanning IDE completions, chat, multi-file agent mode, autonomous cloud coding agent, agentic code review, and Agent HQ — a control plane for third-party agents from Anthropic, OpenAI, Google, Cognition, and xAI, all accessible under the same Copilot subscription.
| Feature | Status (June 2026) | Plans | Source |
|---|---|---|---|
| Inline code completions | GA; unlimited on paid plans | All paid | GitHub Docs |
| Next Edit Suggestions | GA; VS Code, Xcode, Eclipse | All paid | GitHub Docs |
| Copilot Chat | GA; IDE, GitHub.com, Mobile | All paid | GitHub Docs |
| Agent mode (multi-file, autonomous) | GA; VS Code + JetBrains (Mar 2026) | Pro+ and above | GitHub Blog, Feb 2025 |
| Copilot coding agent (cloud) | GA; all paid plans | Pro+, Business, Enterprise | GitHub Blog, May 2025 |
| Agentic code review | GA; Mar 2026 | Business, Enterprise | GitHub Blog |
| Agent HQ (third-party agent orchestration) | Announced; rolling out | Paid Copilot sub | GitHub Blog, Oct 2025 |
| GitHub Spark (natural language app building) | GA; Pro+ and Enterprise | Pro+, Enterprise | GitHub Blog |
| GitHub Copilot desktop app | Technical preview | Pro, Pro+, Business, Enterprise | GitHub Blog, Jun 2026 |
| Custom agents (AGENTS.md) | GA since Universe 2025 | Business, Enterprise | GitHub Docs |
| MCP server integration | GA | All paid | GitHub Docs |
The coding agent — launched May 2025, per the GitHub Blog — is the most strategically significant addition. Assign a GitHub issue to Copilot, and it autonomously researches the repository, creates an implementation plan, commits code to a branch, and opens a pull request for human review. The agent runs on GitHub Actions, which means it uses compute the organization is already paying for, and it requires human approval before any CI/CD workflows run. In terms of agentic capability, it narrows — but does not close — the gap with Claude Code’s terminal-native agentic architecture.
Agent HQ, announced at GitHub Universe 2025, is GitHub’s answer to the fragmentation problem: a single control plane where developers can orchestrate Copilot alongside third-party agents (Claude, Codex, Gemini, Cognition’s Devin) from one unified interface. The strategic logic is sound — GitHub sits at the center of where code lives, and whoever controls the workflow surface controls the agent marketplace. Whether execution delivers on the framing is still being determined.
IDE and Platform Support
| Platform | Features Available | Notes |
|---|---|---|
| VS Code | Full (completions, chat, agent mode, coding agent) | Primary IDE; most features GA here first |
| JetBrains (IntelliJ, PyCharm, WebStorm, etc.) | Completions, chat, agent mode (GA Mar 2026) | Second major family; previously VS Code only |
| Visual Studio | Completions, chat, agent mode | GA 2025 |
| Eclipse | Completions, chat, Next Edit Suggestions | Growing coverage |
| Xcode | Completions, Next Edit Suggestions | Apple-ecosystem support |
| Vim / Neovim | Completions only | No chat |
| GitHub.com | Chat, coding agent, code review | Native web integration |
| GitHub Mobile | Chat | Limited feature set |
| Windows Terminal | Chat | CLI-adjacent |
GitHub Copilot vs Competitors 2026
GitHub Copilot vs Cursor vs Claude Code: Key Metrics
| Metric | GitHub Copilot | Cursor | Claude Code |
|---|---|---|---|
| Paid users / subscribers | 4.7M (Jan 2026) | ~1M+ paying | ~500K est. |
| ARR | $900M–$1.1B (Axis est.) | $2B (Bloomberg, Feb 2026) | $2.5B (Anthropic; Feb 2026) |
| “Most loved” (JetBrains, Jan 2026) | 9% | 19% | 46% |
| Work adoption (JetBrains, Jan 2026) | 29% | 18% | 18% |
| Developer awareness | 76% | 69% | 57% |
| CSAT | Not disclosed | Not disclosed | 91% |
| NPS | Not disclosed | Not disclosed | 54 |
| Primary interface | IDE extension | AI-native IDE (VS Code fork) | Terminal-native agent |
| Base price | $10/mo (Pro) | $20/mo (Pro) | $17/mo (Claude Pro) |
| Fortune 100 presence | 90% | Growing | 75% startups (primary) |
Per JetBrains’ April 2026 AI Pulse analysis, GitHub Copilot’s growth in both awareness and work adoption has stalled since the previous survey wave. Cursor’s growth has also slowed. Claude Code is the only tool in the category showing meaningful acceleration across awareness, adoption, and satisfaction simultaneously — from 3% work adoption in April 2025 to 18% in January 2026, a 6× increase in nine months.
The enterprise picture is more nuanced. In companies with more than 5,000 employees, Copilot’s adoption reaches 40% — significantly higher than its 29% overall — because enterprise procurement, Microsoft 365 bundling, and GitHub Enterprise relationships produce mandatory or near-mandatory deployment at scale. The “two-layer enterprise stack” described by multiple analysts — Copilot for broad autocomplete baseline, Claude Code for senior engineers on complex work — is the most common pattern in large organizations today.
Copilot’s structural advantage is distribution. GitHub hosts over 150 million developers. 80% of new developers on GitHub now use Copilot in their first week, per GitHub Universe 2025. No competitor has access to that onboarding funnel.
GitHub Copilot Developer Survey Statistics
How Developers Actually Use GitHub Copilot
| Metric | Value | Source |
|---|---|---|
| Developers who use AI tools at work (2026) | 84% | Stack Overflow Developer Survey 2025 |
| Developers who use AI tools daily | 51% | Stack Overflow Developer Survey 2025 |
| Developers who trust AI accuracy | 29% | Stack Overflow Developer Survey 2025 |
| Developers who actively distrust AI accuracy | 46% | Stack Overflow Developer Survey 2025 |
| Copilot awareness worldwide | 76% | JetBrains AI Pulse, Jan 2026 |
| Copilot work adoption worldwide | 29% | JetBrains AI Pulse, Jan 2026 |
| Copilot adoption in 5,000+ employee orgs | 40% | JetBrains AI Pulse, Jan 2026 |
| Developers who feel less frustrated with Copilot | 60–75% | Multiple surveys |
| Developers rating Copilot “extremely easy to use” | 43% | Accenture study |
| Developers rating Copilot “extremely useful” | 51% | Accenture study |
| Time to realize productivity gains | ~11 weeks | Microsoft Research |
| Developers saving 1+ hour/week with AI | 90% | JetBrains 2025 |
| Developers saving 8+ hours/week with AI | 20% | JetBrains 2025 |
The 29% trust figure from the Stack Overflow Developer Survey 2025 — down from 40% in 2024 — is the most important trend in developer AI adoption that isn’t being discussed with appropriate seriousness. Developers are using AI coding tools more than ever, while trusting the output less than ever. This isn’t cognitive dissonance; it’s a rational adaptation. You use the tool because it speeds up routine work, and you verify the output because you know it gets complex things wrong.
For Copilot specifically, the trust dynamic intersects with the security statistics: if 29.1% of Python code generated contains potential security weaknesses, and only 29% of developers trust AI output, then the population of developers who trust Copilot-generated Python code without review is a small subset of a small percentage. That’s appropriate behavior, but it means the productivity gains from autocomplete are always partially offset by the review burden that AI-generated code creates.
GitHub Copilot Security Statistics
Security Risks and Safeguards
| Metric | Value | Source |
|---|---|---|
| Code scenarios with security issues (Pearce et al. 2022) | 44% of 89 tested scenarios | Pearce et al. ICSE 2025 proceedings |
| Python code with potential security weaknesses | 29.1% | GitHub research, 2024 |
| Security issue resolution speed with Copilot | 62% faster | Microsoft internal data |
| Copilot-assisted PRs with more security vulnerabilities | 2.74× more than human-only PRs | CodeRabbit analysis |
| Code churn increase (AI-era, 2020–2024) | 3.1% → 5.7% | GitClear, 211M lines |
| Code duplication increase (AI-era) | ~4× increase | GitClear 2024 report |
The 2.74× security vulnerability figure from CodeRabbit’s pull request analysis is not in GitHub’s own research, and it is one of the hardest numbers to square with the 62% faster security issue resolution claim from Microsoft. They can both be true simultaneously: AI-generated code introduces more vulnerabilities during authoring, and Copilot’s code review feature then helps catch them faster in review. Whether the net effect on security posture is positive or negative depends on how rigorously organizations use the review tooling, which is itself an adoption and training question.
The Pearce et al. finding — 44% of scenarios containing security issues across 25 CWE categories — was conducted on an earlier version of Copilot. GitHub has improved its filtering since 2022. But the structural reality is unchanged: a model trained on billions of lines of public code will reproduce the vulnerability patterns that appear in billions of lines of public code. That’s not a failure of this specific model; it’s a consequence of how code generation models work.
Methodology
Axis Intelligence Research collected data from primary sources exclusively: Microsoft SEC EDGAR 8-K and 10-K filings (CIK 0000789019), Microsoft earnings call transcripts, GitHub official blog and documentation, JetBrains AI Pulse survey (January 2026, n=10,000+), InfoQ reporting on peer-reviewed academic studies, and Financial Times reporting with named sources.
CPMI™ Calculation: The Copilot Platform Momentum Index was computed by Axis Intelligence Research using: (a) paid subscriber growth YoY from Microsoft earnings disclosures; (b) enterprise penetration from GitHub and Microsoft official statements; (c) code share from GitHub official telemetry; (d) developer satisfaction from JetBrains AI Pulse “most loved” rating (January 2026, n=10,000+). The satisfaction denominator is multiplied by 1,000 to produce readable index values.
ARR Estimate: The $900M–$1.1B ARR figure is an Axis Intelligence Research calculation, not a Microsoft disclosure. It uses the disclosed 4.7 million paid subscriber count and conservative tier mix assumptions. Actual ARR will differ based on enterprise discount rates and actual tier distribution.
Productivity studies: The 55.8%, 26%, and –19% figures come from distinct studies with different populations and methodologies. They are not comparable in isolation. Axis Intelligence Research presents all three to reflect the actual state of the evidence, not a curated picture.
Data as of: June 25, 2026. GitHub Copilot pricing data as of June 1, 2026 (post usage-based billing transition).
About This Dataset
This dataset covers GitHub Copilot user counts, paid subscriber growth, revenue estimates, market share, pricing tiers, productivity research outcomes, enterprise adoption, and developer survey data for 2022–2026. The CPMI™ metric is an original calculation by Axis Intelligence Research. Dataset updated quarterly.
License: Creative Commons Attribution 4.0 International (CC BY 4.0).
Suggested citation (APA): Axis Intelligence Research. (2026, June 25). GitHub Copilot Statistics 2026: Users, Revenue, Market Share & AI Coding Data. Axis Intelligence. https://axis-intelligence.com/github-copilot-statistics/
Suggested citation (MLA): Axis Intelligence Research. “GitHub Copilot Statistics 2026: Users, Revenue, Market Share & AI Coding Data.” Axis Intelligence, 25 June 2026, axis-intelligence.com/github-copilot-statistics/.
Suggested citation (Chicago): Axis Intelligence Research. “GitHub Copilot Statistics 2026: Users, Revenue, Market Share & AI Coding Data.” Axis Intelligence, June 25, 2026. https://axis-intelligence.com/github-copilot-statistics/.
Cite This Research
<blockquote>
<p>Axis Intelligence Research. (2026). <em>GitHub Copilot Statistics 2026: Users, Revenue, Market Share & AI Coding Data</em>. Axis Intelligence.</p>
<p>Retrieved from <a href="https://axis-intelligence.com/github-copilot-statistics/" rel="dofollow">https://axis-intelligence.com/github-copilot-statistics/</a></p>
</blockquote>
Frequently Asked Questions
How many users does GitHub Copilot have in 2026?
GitHub Copilot reached approximately 20 million total users by July 2025, per Microsoft CEO Satya Nadella’s disclosure in the Q4 FY2025 earnings call. Of those, 4.7 million are paid subscribers as of January 28, 2026, per Microsoft’s FY26 Q2 earnings release. The remaining 15+ million are free users, students, open-source maintainers, and trial accounts.
What is GitHub Copilot’s revenue in 2026?
Microsoft does not publicly disclose GitHub Copilot’s ARR. Axis Intelligence Research estimates $900 million to $1.1 billion for Q2 FY2026, based on the disclosed 4.7 million paid subscribers and public pricing tiers. The Financial Times reported in May 2026 that Copilot was on pace for “at least $550 million annually” — a conservative floor based on the cheapest tier only. GitHub’s total ARR exceeded $3 billion at end 2025, per the same report.
What percentage of code does GitHub Copilot generate?
GitHub Copilot generates an average of 46% of all code written by active users as of 2025–2026, per GitHub’s own telemetry — up from 27% at the product’s commercial launch in 2022. Java developers see the highest rate at 61%. The suggestion acceptance rate (code actually kept by developers) is 27–30%, with 88% of accepted code retained long-term.
How much faster do developers code with GitHub Copilot?
Evidence varies by study. A GitHub/MIT controlled experiment (Peng et al. 2023) found a 55.8% improvement in task completion speed. A larger randomized controlled trial by researchers from Microsoft, MIT, Princeton, and Wharton (2024), covering 4,867 developers at Microsoft, Accenture, and a Fortune 100 firm, found a 26.1% increase in pull requests completed per week. An independent METR study (2025) of experienced developers on complex real-world tasks found a –19% effect. The divergence is explained by task type, developer experience, and the specific metric measured.
What is GitHub Copilot’s market share in 2026?
Among paid AI coding tool users, GitHub Copilot holds approximately 42% market share by subscriber count. By developer work adoption, it is used by 29% of developers worldwide per JetBrains AI Pulse (January 2026), ahead of Claude Code and Cursor, both at 18%. By revenue, GitHub Copilot’s estimated $900M–$1.1B ARR is below both Cursor ($2B ARR) and Claude Code ($2.5B ARR), reflecting those tools’ higher average revenue per user.
What changed with GitHub Copilot pricing in 2026?
On June 1, 2026, GitHub transitioned all Copilot plans from premium request unit-based billing to usage-based billing measured in AI Credits (1 credit = $0.01). Base plan prices did not change: Pro remains $10/month, Pro+ $39/month, Business $19/user/month, Enterprise $39/user/month. Code completions and Next Edit Suggestions remain unlimited on all paid plans. Agentic features, code review, and premium model access now consume AI Credits from each plan’s monthly allowance, with overage billed at per-token rates.
How does GitHub Copilot compare to Claude Code and Cursor?
The three tools serve overlapping but distinct use cases. GitHub Copilot has the largest user base (4.7M paid), the lowest individual price ($10/mo Pro), and the deepest GitHub and enterprise integration. Cursor ($20/mo) is an AI-native IDE with the highest revenue among competitors ($2B ARR). Claude Code ($17/mo via Claude Pro) has the highest developer satisfaction (91% CSAT, 46% “most loved” per JetBrains) and highest ARR ($2.5B). Per JetBrains AI Pulse, the most common enterprise pattern is Copilot deployed broadly as an autocomplete baseline, with Claude Code adopted bottom-up by senior engineers for complex agentic work.
Is GitHub Copilot safe to use for production code?
Research indicates Copilot-generated code carries elevated security risk versus human-written code. A 2022 study by Pearce et al. found security issues in 44% of generated code scenarios across 25 Common Weakness Enumeration categories — though on an older model version. GitHub’s own 2024 research found 29.1% of generated Python code contains potential security weaknesses. CodeRabbit’s pull request analysis found AI-assisted PRs have 2.74× more security vulnerabilities than human-only PRs. GitHub’s mitigations include real-time vulnerability filtering in Copilot’s suggestions and the Copilot code review feature, which can flag issues before merging. Enterprise security best practice is treating Copilot as an accelerator for code authoring, with mandatory human review and automated security scanning on all AI-assisted PRs.
What is the CPMI™ and what does it show for GitHub Copilot?
The Copilot Platform Momentum Index (CPMI™) is an original metric calculated by Axis Intelligence Research: (Paid_Subscriber_Growth_YoY% × Enterprise_Penetration% × Code_Share%) ÷ (Developer_Satisfaction_Score × 1,000). The CPMI™ rose from 4.8 in FY2024 to 34.5 in Q2 FY2026. That increase reflects strong enterprise penetration and code-share metrics — but a satisfaction collapse from ~30% to 9% “most loved” among developers. A rising CPMI™ driven by a shrinking satisfaction denominator is a warning signal, not a success story: it means Copilot’s growth is increasingly dependent on enterprise lock-in rather than product preference.
What are GitHub Copilot’s enterprise plans?
GitHub Copilot Business costs $19 per user per month and includes pooled AI Credits across the organization, IP indemnity, policy controls, audit logs, and SAML SSO. GitHub Copilot Enterprise costs $39 per user per month and requires GitHub Enterprise Cloud (an additional $21/user/month), making the real all-in cost $60 per user per month. Enterprise adds fine-tuned custom models, advanced knowledge indexing, and priority access to new features.